Table.
Characteristics of Medicare Part D Generalist and Oncologist Opioid Prescribers, 2013–2017
| Characteristic | Year | ||||
| 2013 (n = 199 544) | 2014 (n = 204 243) | 2015 (n = 207 857) | 2016 (n = 213 502) | 2017 (n = 212 295) | |
| Prescriber type, No. (% of total) | |||||
| Generalists | 188 343 (94) | 192 823 (94) | 196 180 (94) | 201 537 (94) | 200 049 (94) |
| Oncologists | 11 201 (6) | 11 420 (6) | 11 677 (6) | 11 965 (6) | 12 246 (6) |
| Female, No. (% within specialty) | |||||
| Generalists | 67 773 (36) | 71 045 (37) | 74 010 (38) | 77 704 (39) | 78 423 (39) |
| Oncologists | 3068 (27) | 3229 (28) | 3441 (29) | 3571 (30) | 3736 (31) |
| Medicare patient age, mean (SD) | |||||
| Generalists | 70.7 (4.9) | 70.8 (4.8) | 70.9 (4.7) | 71.1 (4.6) | 71.3 (4.5) |
| Oncologists | 71.3 (3.3) | 71.4 (3.2) | 71.4 (3.2) | 71.6 (3.1) | 71.8 (3.0) |
| Medicare patient risk score, mean (SD)a | |||||
| Generalists | 1.4 (0.5) | 1.4 (0.5) | 1.5 (0.6) | 1.5 (0.6) | 1.5 (0.6) |
| Oncologists | 2.0 (0.4) | 2.0 (0.5) | 2.1 (0.5) | 2.2 (0.5) | 2.2 (0.6) |
| Any opioid prescription, No. (% within specialty) | |||||
| Generalists | 184 364 (98) | 188 056 (98) | 189 809 (97) | 192 952 (96) | 189 753 (95) |
| Oncologists | 11 145 (99.5) | 11 347 (99) | 11 600 (99) | 11 835 (99) | 12 088 (99) |
| Any long-acting opioid prescription, No. (% within specialty) | |||||
| Generalists | 131 464 (70) | 131 206 (68) | 129 746 (66) | 125 977 (63) | 116 497 (58) |
| Oncologists | 10 526 (94) | 10 688 (94) | 10 876 (93) | 11 057 (92) | 11 052 (90) |
Generalist prescribers include family medicine, internal medicine, geriatric medicine, and general practice specialties. Oncologist prescribers include medical oncology, hematology-oncology, and radiation oncology specialties. Physicians included in this table wrote a minimum of 100 Medicare Part D–covered prescription fills for at least 10 Medicare Part D beneficiaries in a given calendar year. All bivariate comparisons between generalists and oncologists were statistically significant in each study year at P < .001 and with absolute standardized mean differences >0.10.
Risk scores reflect the complexity of a prescriber’s Medicare Part D patient population and are calculated using the Centers for Medicare and Medicaid Services hierarchical conditions categories. Risk scores account for patient age, sex, Medicare coverage type, and diagnoses from the previous year.